Version 2 of my checkpoint is a merge of Unholy Desire v8, One Obsession (Anime), Semi-Real MM, Perfection 2.5D, and Spectacular 2.5D.
When deciding which checkpoints to combine, I first create an image with each potential checkpoint. All comparison images are created using the same LoRAs and prompts. Then I compare which lighting, face shape, body shape, color depth, etc., might work well together or not.
Therefore, I decided to proceed similarly with my demo images. I took the prompts from promotional images of the base checkpoints.
Suggested Negative Prompts:
lowres, bad anatomy, (worst quality:1.8), low quality, hands bad, (normal quality:1.3), bad hands, mutated hands and fingers, extra finger, extra legs, duplicate, cropped, jpeg, artifacts, long body, mutated, disfigured, bad proportions, bad feet, ugly, text font ui, missing limb, black bars, ugly hands, extra fingers, logo, patreon username, text, watermark, extra toes,
I use Euler A, 40 Steps, CFG 8, no Refiner, built-in VAE.
